Animals. All animal studies were approved by the Institutional Animal Care and Use Committee at A*STAR. 10 week-old ACI rats (n = 5) were given drinking water containing 100mg/L DEN continuously and followed up to 16 weeks. Control rats (n = 5) were given normal drinking water. During sacrifice, animals were anesthetized and blood was collected by cardiac puncture and death was confirmed by cervical dislocation. Livers were resected, measured by electronic calipers and snap frozen in liquid nitrogen.

MRS. Approximately 48 mg of [1-13C] KIC (Sigma #750832), doped with 15 mM trityl-radical OXO63 and 3 ml of Gd-DOTA (10 mM), was hyperpolarized for 60 min. The sample was subsequently dissolved in a pressurized and heated alkaline solution to yield a HP solution of 80 mM with a polarization of 30%, T1 of 25 seconds at physiological temperature and pH. Rats were positioned in a 9.4 T horizontal bore MR scanner interfaced to a Avance III console, and inserted into a dual-tuned (1H/13C) rat abdominal coil. Anatomical positioning was confirmed by the acquisition of a coronal FLASH image (TE/TR, 8.0/100.0 ms; matrix size, 192x192; FOV, 50 x 36mm; slice thickness, 2.0 mm; excitation flip angle, 30°). Upon intravenous injection (0.5mmol/kg BW), a respiratory-gated 13C Single-Pulse MRS acquisition was initiated. Thirty individual liver spectra were acquired over 1 min (TR, 2 s; flip angle, 25°; SW, 8 kHz; acquired points, 2,048). Liver 13C MR spectra were analyzed using the AMARES algorithm as implemented in the jMRUI software package. Spectra were baseline and DC offset-corrected based on the last half of acquired points. To quantify hepatic metabolism, the spectra were summed over the first 30 s upon 2-ketoisocaproate arrival. Metabolite peaks corresponding to [1-13C]2-ketoisocaproate (172.6ppm) and its metabolic derivatives [1-13C]leucine (176.8ppm) and [1-13C]bicarbonate (160.8 ppm) were fitted with prior knowledge assuming a Lorentzian line shape, peak frequencies, relative phases, and linewidths. For each animal, tCarbon is defined as the sum of all these three metabolite peaks. The normalized ratios [1-13C]leucine/tCarbon and [1-13C]bicarbonate/tCarbon were computed for statistical analysis.

16 weeks upon treatment with DEN, rats developed liver cancer and multiple tumors began to develop, as displayed in the coronal MRI images (Figure 1A). The metabolic fate of [1-13C]KIC illustrated the catabolic function of BCKDH (Figure 1B). Probing of BCKDH enzyme activity in vivo with hyperpolarized [1-13C]KIC reveals a significant reduction in bicarbonate production between control and DEN animals (p < 0.05) (Fig. 1C & D). This result corroborated with ex-vivo metabolomics and biochemical assays analysis. No significant change in leucine production was observed.

Previously, Karlsson et al investigated the transamination of hyperpolarized [1-13C]KIC into [1-13C]leucine mediated via the branched chain amino acid transferase (BCAT) in rodent models of murine lymphoma and rat mammary adenocarcinoma, and they found significant increase in leucine production compared to the respective surrounding intestine and muscle tissues (2). This comparison is inaccurate because these control tissues are of different phenotypes from the tumors and therefore possess distinct metabolic profiles from one another. Perhaps limited by the low magnetic field strength of 2.35T, no detection of [1-13C]bicarbonate was observed in any of these models. In contrary, we compared in vivo BCAA metabolism in healthy control and HCC livers at 9.4 T and successfully detected both leucine and bicarbonate. To our knowledge, this is the first published report of probing liver cancer BCKDH activity in vivo with hyperpolarized carbon-13 MRS.